Hello John
Thanks for your reply
I have a question about the image distortion. How to determine the
extent of a image is distorted?
Do you mean that If the distortion is modest the coregistion is necessary?
And how to normalise the functional images without coregistion?
thanks
On 6/22/05, John Ashburner <[log in to unmask]> wrote:
>
> > I have some questions about preprocessing of fMRI data sets. My
> > preprocessing steps are :
> >
> > 1. realignment of the functional volumes;
> > 2. coregister of the mean functional volume and the structure image(T1);
> > 3. Normalise functional image with T1 template and T1 structure image.
> > 4. smoothing the result of normalise with a FWHM 2-3 times of the voxel
> > size.
> >
> > Are these steps correct?
> >
> > Some users don't coregister between the functional image and structure
> > image. And they normalise the functional images with a EPI.mnc. Which
> > method is better?
>
> This question can only be answered empirically. Whatever works better will
> depend on your images. This includes whether their contrasts are similar to
> the SPM templates or not, and also the artifacts they contain. For example,
> if your EPI images are distorted, then you won't achieve a good registration
> between the anatomical and functional images. If this is the case then
> applying warps estimated from the anatomical images to the functional images
> is unlikely to work well.
>
> Any method or model is only valid with respect to the data that it is to be
> applied to. This validity will depend on whether the model assumptions hold
> for that data.
>
> Best regards,
> -John
>
>
|