I am conducting a longitudinal VBM study to investigate GM density
response to treatment in a depressed patient population using DARTEL
processing. I have two images for each subject, pre and post
treatment. I am interested in conducting within-group and
between-group comparisons.
I have previously been advised to create one difference image for each
subject by subtracting their baseline warped and modulated (mwc1*)
image from their follow-up image using SPM5 ImCalc, I have no problem
with this step. I have then been advised to spatially normalize the
difference image to MNI space, and then smooth the image, and conduct
statistical analyses using the difference images.
My issue: I have an individual flow field (u rc1*.nii) for each
subject's image at baseline and follow-up which I can use to backward
map from the template to individual (the first step in the
normalization to MNI space process). I do not have a flow field for
the difference images. How can I normalize the difference images to
MNI space without a flow field specific to the difference image?
Can I simply take the baseline and follow-up images which have
undergone segmentation with DARTEL, normalize them to MNI space and
then create difference images?
I appreciate any help anyone can offer.
Thanks,
Jennifer Addicks
Graduate Student
Department of Neuroscience, University of Ottawa
Mood Disorders Research Unit, University of Ottawa Institute of Mental
Health Research
1145 Carling Ave, Research Tower, Room 6412
Ottawa, Ontario K1Z 7K4
Tel: 613.722.6521 x.6188
Fax: 613.798.2982
[log in to unmask]
www.imhr.ca
On Tue, Dec 16, 2008 at 6:09 PM, Jennifer Addicks
<[log in to unmask]> wrote:
> Hello,
> I am conducting a longitudinal VBM study using SPM5. I am interested
> in examining GM density changes in depressed patients pre and post
> treatment. Specifically, I have defined 2 groups of patients based on
> treatment outcome (remitters and nonremitters) and I would like to
> conduct within and between group comparisons of GM density changes. I
> have successfully completed my preprocessing of all patients' baseline
> and followup scans using DARTEL, and I now have some questions
> concerning my 2nd level factorial design specification and contrasts.
>
> 1) I have set up a Full Factorial Design as follows:
> Factor 1: Outcome Group levels: 1-nonremitters; 2-remitters
> independence-yes, variance unequal
> Factor 2: Time levels: 1-baseline; 2-followup independence-no,
> variance unequal
> Covariates: Age(baseline), Age(followup), gender
> Question: To compare for ex. GM density change within the nonremitter
> group over time, can I simply define a t contrast of
> [-1 1 0 0 0 0 0] ? Because I have specified that the time factor is
> not independent does this make this a paired t-test or is the analysis
> treating this group as 2 groups rather than one group at 2 time
> points? Is there a way to mask one contrast with another to examine
> the differences in atrophy/growth seen in the 2 groups over time?
>
> 2) I have also tried to set up a straight-forward paired t-test
> analysis by entering patients' baseline and followup scans as pairs. I
> have some questions if this is a more appropriate analysis:
> First, is it possible to add covariates to this analysis- for example
> scan duration which differs among subjects? How would the contrast be
> set up for this analysis?
>
> 3) Finally, I want to perform a multiple-regression to investigate the
> role of certain predictor variables (level of treatment resistance,
> aggressiveness of treatment) on GM density change in patients (whole
> group together). I would like to use GM density difference images
> (representing areas of GM density change between baseline and
> followup)- do I already have such images from the DARTEL
> preprocessing? If not, how would I create such images? How would I set
> up this contrast?
>
> Thank you,
>
> Jennifer Addicks
> Graduate Student
> Department of Neuroscience, University of Ottawa
> Mood Disorders Research Unit, University of Ottawa Institute of Mental
> Health Research
> 1145 Carling Ave, Research Tower, Room 6412
> Ottawa, Ontario K1Z 7K4
> Tel: 613.722.6521 x.6188
> Fax: 613.798.2982
> [log in to unmask]
> www.imhr.ca
>
|