George,
I was looking for an answer to a similar question but have not yet got one.
This is what I think your contrasts mean (assuming interictal is cond. 0
& ictal cond. 1):
[-1 1 -1 1] Interictal < Ictal Independent of Group
[1 -1 1 -1] Interictal > Ictal " "
[1 -1 -1 1] (Group 1: Interictal > Ictal) & (Group 2: Interictal < Ictal)
[-1 1 1 -1] (Group 1: Interictal < Ictal) & (Group 2: Interictal > Ictal)
If my interpretations are right then I would have thought
[1 1 -1 -1] Group 1 > Group 2 Independent of Condition
but this contrast is not accepted. I don't know why or what I should do
to see this.
Joel
-----------------------------------------------
Joel Dunn
Research Assistant
PET Imaging Centre
Guy's, King's and St Thomas' School of Medicine
Lower Ground Floor, Lambeth Wing,
St Thomas' Hospital, Lambeth Palace Road,
London SE1 7EH UK Tel: 020 7188 7445
Email: [log in to unmask] Fax: 020 7620 0790
-----------------------------------------------
George wrote:
> Dear SPMers,
>
>
>
> I am looking changes in blood flow in patients with epilepsy using
> SPECT scans. I have a "during seizure" (ictal) state and an
> "in-between" seizure (interictal) state for each patient. These
> patients have been grouped according to their time of "injection
> before seizure ends" and "after seizure ends". Within the groups I
> want to look at a simple subtraction of increases vs. decreases of
> blood flow. However, I also want to look at the differences between
> groups. Any ideas on how I would set up my model.
>
>
>
> Currently, I am setting up my model, using multigroup: conditions and
> covariates with 2 groups (Group 1 = Early injection, Group 2= Late
> injection), # of subjects within group (n=18) and I select the
> interictal and ictal scans for each subject and set the condition as 0
> 1. I repeat the same thing for the second group.
>
>
>
> I set the covariates and nuisance variables at 0, proportionally
> scaling the global mean to 50, set relative threshold masking at 0.8,
> set global calculation to mean voxel value, and do not apply a
> non-sphericity correction.
>
>
>
> I have set up four different contrasts to look at my differences
> within and between groups.
>
>
>
> To see Early Injection (Group1) > Late Injection (Group 2) and Ictal >
> Interictal, I set my contrast at [-1 1 -1 1].
>
> For Group 1 > Group 2 and Interictal > Ictal, I set my contrast at [-1
> 1 1 -1]
>
> For Group 2 > Group 1 and Ictal > Interictal, I set my contrast at [1
> -1 -1 1]
>
> For Group 2 > Group 1 and Interictal > Ictal, I set my contrast at [1
> -1 1 -1]
>
>
>
> The results I get seem like they would be the correct results, but I
> am unsure at how SPM calculates these results. Any ideas would be
> extremely helpful. Thank you very much for you time,
>
>
>
> Sincerely,
>
>
>
> George Varghese
>
>
>
>
>
> George Varghese
>
> Research Assistant
>
> Yale University
>
> Department of Neurology
>
> New Haven, CT 06510
>
> phone: (203) 785-6411
>
> fax: (203) 737-2538
>
> email: [log in to unmask]
>
>
>
|