Hi,
You four steps look perfectly good to me, so I cannot understand why you see this shift upwards. Can you post an illustration of the problem (load a jpg/png/gif onto a third party website and link to it, as the email list doesn't accept anything except very small attachments)?
It doesn't matter much whether you use brain extracted or non-extracted functional scans, as epi_reg looks at the grey-matter/white-matter border, so as long as that is present it doesn't care about the intensities outside of the brain. This is not true for the fieldmap though - as you do need both extracted and non-extracted versions of the fieldmap magnitude scan (if you are using that).
All the best,
Mark
On 5 Sep 2013, at 21:01, Kyle Heidtman <[log in to unmask]> wrote:
> Hi
>
> I have some partial FoV functional scans that I'm trying to register to standard space. I don't have any functional images of the entire head. However, I do have T1 scans that were taken at the same time as the functional scans.
>
> I've looked on the FSL wiki and done some experimenting. I've found that epi_reg seems to be fairly robust against partial FoV effects, so I've been taking the following steps:
>
> 1. I use epi_reg to to register the functional images to T1 space.
>
> 2. I use flirt to register the t1 scan to the MNI standard space image.
>
> 3. I concatenate the functional-to-T1 transformation matrix with the T1-to-standard transformation matrix.
>
> 4. I apply the concatenated matrix to the original functional scan, registering it to standard space.
>
> Is that how I am supposed to do that? Or is there a better way?
>
> However, I've run into a strange problem. I've compared the image output by epi_reg to the image being used as a reference volume, and the output image always looks like it's been shifted a few pixels upwards (in the superior direction).
>
> I don't know what could be causing this. Do any of you have ideas?
>
>
> Also, I have a few other questions about epi_reg
>
> Am I supposed to input skull-stripped functional images or whole head functional images to epi_reg?
>
> If I am supposed to submit whole head functional images, can I skull-strip the scans after they've been registered to standard space?
>
>
> Thank you for your time,
> Kyle
|