Thanks Marko!
It does make sense to coregister the images to each other first... Thank you for explaining that.
Just to verify that I understand, here is the plan:
1) Segment Time1 and Time2 in native space.
2) Coregister Time2 segments to Time1 segments (estimate only).
3) Apply the sn.mat from Time1 -> MNI to Time1 images and "preserve amount" to modulate.
4) Apply the .mat from the Time1->Time2 coreg and the Time1->MNI sn.mat to Time2 images (also selecting "preserve amount").
5) Mask the images to get rid of values < .1
6) Smooth
7) Subtract the MNI-space Time2 segments minus MNI-space Time1 segments.
Is this right? Is there a simple way to combine .mat files for step 4 so the images will not have to be resampled twice?
Also, when I segment Time1 in native space, then use Normalize to apply the sn.mat and "preserve amount", should I be concerned that I get a different result than when Segment directly produces modulated, normalized segments? When I subtract these images I was hoping to get all 0, but have values ranging from about -.2 to .2.
Thanks again,
Dana
-----Original Message-----
From: Marko Wilke [mailto:[log in to unmask]]
Sent: Tuesday, August 12, 2008 3:56 AM
To: Dana Perantie
Subject: Re: [SPM] longitudinal VBM
Hi Dana,
the approach in general sounds good, but you should make sure that the
difference between the images stems from the images and not from the
normalization. In other words, it certainly makes sense to use the same
set of normalization parameters for each pair of images (coregister
before applying the parameters).
Also, this being kids, you may want to make sure that your processing is
appropriate for this population (it is something of a ritual to always
say that :)
Best,
Marko
Dana Perantie wrote:
>
>
> Dear SPM list,
>
> I am planning analyses of two groups (patients and controls) with
> structural MRI at baseline and after 2-years. The age range is 7-16 at
> baseline. We want to compare change over time between the two groups.
> I was wondering if it would be OK to calculate change images by
> subtracting (Time2 minus Time1 modulated, smoothed images) and use these
> images in a 2-sample t-test. Would mask out voxels < .1 on any image at
> either time point.
>
> Within the patient group, I want to test correlations between brain
> change and clinical variables. With change images, I could use the
> multiple regression model.
>
> Can someone please tell me if this makes sense - or what would be wrong
> with it and what would be preferable?
>
> Thank you,
>
> Dana
>
|