Griselda:
1 -1 -1 1 is correct.
You should also do the inverse: -1 1 1 -1
Best,
Darin Dougherty, MD
Massachusetts General Hospital
> -----Original Message-----
> From: Griselda Garrido [SMTP:[log in to unmask]]
> Sent: Thursday, April 25, 2002 2:14 PM
> To: [log in to unmask]
> Subject: 2x2 factorial design
>
> Dear SPMerīs
>
> We are working on a SPECT dataset with one group of patients (group 1,
> n=9) and one group of healthy controls (group 2, n=10), both of which
> studied during a cognitive activation task (one scan per subject) and a
> baseline condition (one scan per subject).
>
> We wish to investigate whether there are brain regions in the group of
> patients that show greater activation than controls during the
> activation task relative to the baseline condition. We believe that a
> 2x2 factorial design would apply to such question, but we are uncertain
> about the best way to design our SPM contrasts. This is what we tried:
>
> Multi-group:conditions & covariates
> #groupīs: 2
>
> group 1 # subjects: 9
> [2] group 1: subject 1: condition? ab
> ...
> [2] group 1: subject 9: condition? ab
>
> group 2 # subjects: 10
> [2] group 2: subject 1: condition? ab
> ...
> [2] group 2: subject 10: condition? ab
>
> We would like to test this:
> (Patients Activation - Patients Baseline) - (Controls Activation -
> Controls Baseline)
>
> We have searched the list but we are still uncertain if the contrast
> 1 -1 -1 1 give us what we want?
>
> We appreciate your help.
>
> Griselda J Garrido
> University of Sao Paulo - Brazil
> www.directnet.com.br/users/griselda
|