Hello Stephen and Mark,
Thank you very much for the quick response. I'm sorry I wasn't very clear in describing what I was trying to do. I am doing the typical registration sequence (EPI to T1-weighted structural using BBR, and structural to standard using FNIRT). The EPI-to-structural step is working fine, but the FNIRT normalization of my structural is not. This seems to be particular to FNIRT, because if I do linear registration with 12 DOF it works fine. I have noticed in my structural, before skull-stripping, there is a large signal from fat. Is it possible that that would interfere with the nonlinear registration? I've attached a picture of the normalized and raw anatomical for reference.
Thanks again for any advice you may have.
-Ian
On Mar 23, 2013, at 12:28 AM, Mark Jenkinson wrote:
> Dear Ian,
>
> If I am understanding your email correctly, you are trying to use FNIRT to register the EPI to the structural. This is not what we recommend, for all sorts of reasons related to image contrast, resolution, distortions and dropouts. Instead we recommend using FNIRT for the structural to standard registration and FLIRT with BBR for the EPI to structural registration, and then combine these registrations. See our documentation and the FSL Course for more details on this.
>
> All the best,
> Mark
>
>
> On 23 Mar 2013, at 01:46, Ian Ballard <[log in to unmask]> wrote:
>
>> Hello,
>>
>> I am trying to normalize my EPIs using the FNIRT tool and am getting really
>> dramatic distortions in some of the subjects. The brain is really
>> smeared out and caved in. However, FLIRT runs fine. The
>> bottleneck seems to occur in the highres to standard to step. I am using
>> all of the defaults. I've attached a copy of the log file. I would really
>> appreciate any advice anyone might have.
>>
>> Thank you,
>>
>> Ian
>> <highres_head_to_standard_head.log>
|