Print

Print
Hi Anderson,

thank you for the response. Indeed, on my first level, I have the 2 task conditions (X and Y) as my only EVs, and then in the group level I treated the groups as the EVs, and have the contrasts (Group A > Group B, Group B > Group A...), and then looked at the COPEs defined on the first level of the only the conditions themselves (X or Y)  to see the group difference on one condition.

Sorry, but I just had a quick follow up. For another study dealing with 2 groups again and a simple rest.A.rest.A.... design, on the first level, one would simply define the A as the lone EV, and never rest? And then to examine group differences between the condition A, simply set up the groups as the EVs on the higher level analysis, set the contrasts to Group 1 > Group 2, Group 2 > 1? Sorry for the simply questions, since I know these designs can be easily handled through the model wizard, but I just thought I'd ask to be sure.

Thank you again.

Paul

On Sun, Jan 25, 2015 at 11:29 AM, Anderson M. Winkler <[log in to unmask]> wrote:
Hi Paul,

It appears to be right. However, without seeing the actual design matrices, it makes it difficult to comment. I guess in the 1st level, the EVs are X and Y only, and in this order, and that in the 2nd level the EVs are one for each group, and that you'd apply it to the correct COPEs brought from the lower level. If that's the case, then it should be right.

All the best,

Anderson


2015-01-24 6:09 GMT+00:00 Paul Dhami <[log in to unmask]>:
Greetings FSL experts,

I feel like this is a simple question that has been answered more than once, but one I want to be sure about it...

I have a study with 2 groups (A and B), who were scanned once under the same 2 conditions (X and Y). If I ultimately wanted to contrast the group differences for the conditions (specifically one condition such as Y), would the contrasts from first level to high level be as follows?:

Level 1:

For each subject, avoid any direct contrasts between conditions so...

cope 1 : [1 0] (condition X )

cope 2: [0 1] (condition  Y)

Then for my higher level analysis:

Set up a contras which looks at group differences as in:

Group A > B [1 -1]

Group B > A [-1 1]

and then from here, look at the lower level contrast of condition X and Y, to view the group differences in relation to each contrast?

Any input on this would be greatly appreciated. Thanks again.