 |
 |
Jef, You should also have columns for the interaction of time and condition, so I would correct that first. Then I would look at my older posts on generating contrasts. The contrast generation starts with the null hypothesis. In your example, you are testing if pre and post are different. The contrast - if the columns are described correctly: 1 -1 0 0 zeros(1,17) because you are testing if pre is different from post and pre and post are the first 2 columns. When adding the interaction term to the model, the contrast should become: 1 -1 0 0 1/2 1/2 -1/2 -1/2 zeros(1,17). Hope this helps. Best Regards, Donald McLaren, PhD On Tue, Jul 21, 2015 at 1:41 PM, Jeffrey West <[log in to unmask]> wrote: > Hello > > I am working with spm 8 and a 17 subject one group, 2 time(pre post) X 2 > condition (on off) ANOVA for fmri data. I currently have an estimated model > which has column 1 pre, column 2 post, column 3 on, column 4 off; then the > 17 subjects complete the design matrix. > > This is a new design for me; I am use to the design matrix displaying the > subjects before the conditions. I have made several attempts as to get the > subjects first in the design matrix and have been unable to do so. > > I then moved on to make the contrast but since the model is arranged in a > different format my contrast (ie: main effect of time) does not work. > > Interestingly when I use the contrast for main effect of time as > zeros(1,17) 1 1 -1 -1 it did in fact accept the contrast but the weight was > listed at column 18; thus over the subjects not over the conditions and > because of this I do not believe the contrast is showing the activations > associated with the main effect of time. > > Could someone please with some ANOVA experience in spm8 please help me > understand why the subject are listed after the conditions; even though I > entered the subjects as factor 1 and main effect 1 in the flexible > factorial ANOVA. > > And please help me to understand how to adjust the contrasts accordingly > to match up with the model or is the contrast I used correct for the main > effect of time. > > thank you for any help with this matter. > > Jef > > > > Jeffrey West, M.A. > Research Analyst > Maryland Psychiatric Research Center > Baltimore, Maryland 21228-0247 > Phone: 410-402-6018 > email: [log in to unmask] > >