Hi,
This sounds like you are trying to use fnirt to register a functional
(EPI) volume
to the structural. Is that the case? We do not recommend this.
Instead you
should register the EPI to the structural using a rigid-body
(within-subject)
registration using flirt, ideally with fieldmap correction using fugue.
This is
all done as part of the registration and pre-processing in FEAT if you are
doing fmri. If the image is a diffusion EPI instead then the same
applies,
even the pre-processing with fugue, which you can run in FEAT from the
pre-stats tab.
All the best,
Mark
Vina Goghari wrote:
> Hi Mark,
> Yes part of the bottom of the brain is not covered (3.125 voxel & 26
> slices, 1.5 TR). I also used a full brain search, if that helps. Just
> wondering if it is inappropriate to use non-linear in this case?
>
> Thanks,
> Vina
>
> Mark Jenkinson wrote:
>> Hi,
>>
>> Are you using a partial field of view?
>> Is the cut-off due to it being at the bottom of the image in either the
>> input or reference image?
>>
>> All the best,
>> Mark
>>
>>
>> On 12 Nov 2008, at 13:54, Vina Goghari wrote:
>>
>>> Thanks that explains part of it. With the linear transformation
>>> there is part of the temporal cut-off (which appears to be
>>> appropriate). With the non-linear it appears the brain is stretched,
>>> there appear to be parts of the temporal lobe present that was not
>>> present before.
>>>
>>> Thanks,
>>> Vina
>>>
>>> Steve Smith wrote:
>>>> Hi - we definitely see improved registrations (of the structurals
>>>> to standard space) when adding FNIRT on top of the affine
>>>> FLIRT-derived transformation. One possibly confusing factor is the
>>>> end-slice interpolation is currently different when using FNIRT
>>>> (that'll be made more consistent in future) so it may appear that
>>>> the very top and bottom of the data are 'stretched' (though this
>>>> shouldn't affect any higher-level analyses as it should be outside
>>>> the standard space brain space). [Jesper - did I get that bit right?]
>>>>
>>>> Cheers.
>>>>
>>>>
>>>>
>>>> On 12 Nov 2008, at 01:53, Vina Goghari wrote:
>>>>
>>>>> Hi,
>>>>> I used the older FSL version previously with FLIRT to register my
>>>>> images. With this new version I also used the nonlinear Warp
>>>>> Resolution as well with default of 10 (same subject). The other
>>>>> search options remained the same. However, it appears the
>>>>> registration does not look as good with the new version in some
>>>>> regards? I wasn't able to attach images...
>>>>>
>>>>> Thanks for any insight you can provide!
>>>>>
>>>>> Vina
>>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------------
>>>>
>>>> Stephen M. Smith, Professor of Biomedical Engineering
>>>> Associate Director, Oxford University FMRIB Centre
>>>>
>>>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>>>> +44 (0) 1865 222726 (fax 222717)
>>>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>>>> ---------------------------------------------------------------------------
>>>>
>>>
>
|